Online text statistics: characters (with/without spaces), whitespace, UTF-8 bytes, CJK/Latin word breakdown, lines, non-empty lines, paragraphs, and reading time.
TL;DR: Free online text statistics tool for characters (with/without spaces), whitespace, UTF-8 bytes, CJK/Latin word breakdown, lines, non-empty lines, paragraphs, and reading time. Supports mixed East Asian and Latin text, runs locally in the browser.
For local tools, open DevTools → Network and search your input — it should not appear in requests.
Usage Guide
Word count is rarely just “how many words”: essays hit limits, translators bill by word, and copywriters care about reading time. This page reports characters (with and without whitespace), whitespace count, UTF-8 bytes, CJK and Latin breakdowns plus total words, lines and non-empty lines, sentences, paragraphs, and an estimated reading time — all computed locally from what you paste.
1. Count a draft
Publisher portals disagree on dictionaries. Treat this panel as a planning instrument and reconcile against the official counter before a hard deadline. Save a screenshot of both counters when disputes are likely.
- Paste the full manuscript or the section you are trimming.
- Map each metric to the platform rule you must satisfy (spaces counted or not; CJK by character vs Latin by tokens).
- Edit, then recheck reading time and the metric your editor actually enforces.
2. CJK versus Latin “words”
Following common office intuition, each Han, kana, or Hangul character counts as one word here, while Latin and digit runs count as whitespace-separated tokens. That differs from naive space-only splitting and is closer to how Word or WPS treat Chinese. Mixed documents should be read via the CJK and Latin subtotals, not the single total alone.
- Check subtotals on bilingual paragraphs.
- Do not force English whitespace rules onto Chinese copy.
- Agree the counting rule with the other party in writing for paid work.
3. Metrics people overlook
UTF-8 byte length matters for API payloads and SMS budgets. Non-empty line count matters for checklists and poetry line limits. Reading time assumes roughly 400 CJK characters per minute and 200 Latin words per minute — a planning signal, not a lab measurement. Sentences and paragraphs help when a brief says “keep it under N paragraphs” rather than a word cap.
4. Whitespace and invisible characters
Characters-with-spaces versus without can diverge a lot. Pastes from the web may carry non-breaking spaces or zero-width characters that inflate counts. Run a whitespace cleaner first when this counter and your editor disagree by more than a few units on a short sample. Prefer plain-text paste over rich HTML when submitting to forms that strip markup unpredictably.
5. Combine with replace and dedupe
While trimming to a limit, pair this counter with find/replace and line dedupe: remove filler, collapse duplicate bullets, recount. Keeping the loop in the browser avoids shipping drafts to third-party counters that retain text. For list-heavy documents, non-empty lines often track progress better than word totals.
6. Acceptance checklist
Ship only when your recorded metric matches the form field name — “words” on a Chinese portal may mean characters. When a client disputes a translation word count, export both parties’ counters on the same plain-text sample so you argue about rules, not about different invisible characters in two pastes.
- Compare one paragraph against the target editor counter; small deltas from rule differences are normal.
- If deltas are huge, strip HTML tags and odd whitespace, then recount.
- Record which metric (words vs characters vs bytes) the submission form actually uses.
7. What this page does not claim
It does not implement every CMS proprietary dictionary, does not strip markdown syntax unless you paste rendered text, and does not bill for you. Use it to steer drafting; let the billing or submission system be the final authority.
8. Billing and editorial disputes
Translation and content vendors often disagree because one party counts Word’s “words” while another counts characters without spaces. Before work starts, write the metric name, tool, and language mode into the statement of work. During delivery, both sides should paste the same plain-text export into this counter and into the contractual tool, then photograph both results. If numbers diverge, hunt invisible characters and HTML residue before arguing about rates. For bilingual marketing pages, report CJK and Latin subtotals separately so neither side can claim the total was inflated by the other script. Reading time is for planning briefs and landing-page pacing — never use it as a billing unit. Pasting Markdown counts fences and syntax as characters — fine for raw drafts, misleading for visible words. Prefer rendered plain text when the brief cares about reader-facing length. Code samples inflate Latin word counts; exclude appendix listings from the billed section when contracts say so.
- Name the metric in the contract before drafting starts.
- Compare identical plain-text pastes when disputing counts.
- Report CJK and Latin subtotals on bilingual deliverables.
FAQ
Neither. Counting happens entirely in your own browser — the text is never uploaded and nothing is kept on file.
Real-time counts of characters (with/without spaces), whitespace, UTF-8 bytes, words (CJK + Latin combined), CJK characters, Latin/digit words, lines, non-empty lines, sentences, paragraphs, and estimated reading time — all for mixed East Asian and Latin text.
It follows Word/WPS conventions: each Han character (plus kana and Hangul syllables) counts as one word, while Latin and digit runs split on spaces; the UI lists CJK characters and Latin words separately. Character counts use Unicode code points (one emoji = one character), and UTF-8 bytes are reported alongside for API length checks.
UTF-8 bytes — API and database limits are usually byte-based; non-empty lines — a closer proxy for real list items; reading time — estimated at roughly 400 characters per minute for Chinese and 200 words per minute for Latin text, a planning signal rather than a lab measurement.
Counting rules differ between tools, so small gaps are normal; defer to the target platform's rules before submission. If the gap is huge, check first for hidden whitespace or HTML tags copied from a web page — run a whitespace cleanup before counting again.
While trimming to a limit, watch the counters update as you edit; for lists, deduplicate first and then read the non-empty line count. The tools chain together in one workflow.
Related Tools
Related Articles
Online Word Counter: Count Words, Characters, Lines Instantly
Learn how to use an online word counter for words, characters, lines, and paragraphs. Understand CJK vs Latin counting, reading-time estimates, and writing/dev use cases.
Markdown to Word: Native Style Mapping and the Inline-Format Downgrade List
Turning a Markdown draft into a deliverable .docx: pandoc CLI vs in-browser local conversion, how headings/tables/code blocks map to native Word styles, which inline formats degrade to plain text, and why 'editable' beats 'looks right'.
Word to Markdown: Two Clean Paths and What Survives the Conversion
How to convert .docx to Markdown cleanly: pandoc CLI vs in-browser local conversion, a format fidelity table for headings/tables/images/auto-numbering, and fixes for the .doc legacy format, image bloat, and merged-cell tables.